Improving the accuracy of operation coding in ...

Downloaded from http://qir.bmj.com/ on September 23, 2015 - Published by group.bmj.com

BMJ Quality Improvement Reports 2014; u202053.w1990 doi: 10.1136/bmjquality.u202053.w1990

Improving the accuracy of operation coding in surgical discharge summaries Eirini Martinou, Genevieve Shouls, Nadine Betambeau St George's Healthcare NHS Trust

Abstract Procedural coding in surgical discharge summaries is extremely important; as well as communicating to healthcare staff which procedures have been performed, it also provides information that is used by the hospital’s coding department. The OPCS code (Office of Population, Censuses and Surveys Classification of Surgical Operations and Procedures) is used to generate the tariff that allows the hospital to be reimbursed for the procedure. We felt that the OPCS coding on discharge summaries was often incorrect within our breast and endocrine surgery department. A baseline measurement over two months demonstrated that 32% of operations had been incorrectly coded, resulting in an incorrect tariff being applied and an estimated loss to the Trust of £17,000. We developed a simple but specific OPCS coding table in collaboration with the clinical coding team and breast surgeons that summarised all operations performed within our department. This table was disseminated across the team, specifically to the junior doctors who most frequently complete the discharge summaries. Re-audit showed 100% of operations were accurately coded, demonstrating the effectiveness of the coding table. We suggest that specifically designed coding tables be introduced across each surgical department to ensure accurate OPCS codes are used to produce better quality surgical discharge summaries and to ensure correct reimbursement to the Trust.

operation or procedure that the patient has undergone.

Problem Upon patient discharge from the breast and endocrine surgery unit at St George’s Hospital, discharge summaries are issued to the patient, the GP, and the hospital’s coding department. Our surgical department had received complaints from patients and the hospital coding team that there were discrepancies on our surgical discharge summaries; the type of operation documented in the OPCS coding section was different to the operation that was described in the free text box of the discharge summary. In addition, the junior doctors that complete the majority of discharge summaries were dissatisfied with the lack of specific training regarding OPCS coding. Consulting the online database of OPCS codes was often difficult and time consuming, and a "best guess" was used to decide the most appropriate OPCS code for a particular operation. This was further complicated by the use of abbreviations by surgeons in the patient’s operation note. Accurate documentation is important and any conflicting information is confusing for the GP and other healthcare professionals and can increase patient anxiety. Coding errors may also have financial implications for the Trust if errors lead to reimbursement underpayment.

Background Providing an accurate hospital discharge summary is an important responsibility of a doctor as set out by the General Medical Council.[1] The OPCS coding section is an essential part of all surgical discharge summaries since the OPCS code relates to the

NHS operations are coded using the Office of Population, Censuses and Surveys Classification of Surgical Operations and Procedures (OPCS) system. These operation codes, together with diagnostic codes, are converted to generate Healthcare Resource Group (HRG) codes. It is these HRG codes that are used by “payment by results”[2] to generate the tariff for reimbursement to the Trust. Clinical coding is complex and prone to inaccuracy. Several hospital departments have reported coding errors in up to 47% of cases.[3, 4] Coding errors may cause significant financial losses to departments and Trusts, so any methods used to improve the accuracy of OPCS coding should be encouraged.

Baseline measurement Breast and endocrine surgical discharge summaries were retrospectively collected and analysed from electronic records and patient clinical notes over a two month period. The electronic free text description of the operation was cross referenced with the written operation note. The correct OPCS code for the operation performed was identified using the OPCS handbook provided by the hospital’s coding department. The correct OPCS code was then compared to the code given on the discharge summary and any discrepancies noted. Ninety-two operations were performed in the breast and endocrine unit during the two month study period. Of these, 82% were breast procedures, 12% were endocrine procedures, with minor day-case surgical procedures accounting for the remaining 6% (eg lipoma

Page 1 of 3 © 2014, Published by the BMJ Publishing Group Limited. For permission to use (where not already granted under a licence) please go to http://group.bmj.com/group/rights-licensing/permissions.

Downloaded from http://qir.bmj.com/ on September 23, 2015 - Published by group.bmj.com

excision). Results showed that 32% of operations performed had been inaccurately coded. Calculation of any potential tariff difference between the documented and the correct OPCS code demonstrated that for most inaccuracies, the OPCS code used on the discharge summary underestimated the actual cost of the procedure (table 1). The estimated potential financial loss within our department over the twomonth period was calculated to be £17,000. See supplementary file: ds3257.jpg - “Tariff table”

Design The aim of the study was to increase the accuracy of OPCS coding used on discharge summaries and to ensure consistency of the documented information. A working group was formed within the breast and endocrine department consisting of a consultant surgeon, a surgical senior house officer, and a surgical foundation doctor. The team analysed the problematic areas using the PDSA methodology and liaised with the coding and financial department of the Trust. This clinical and non-clinical multidisciplinary group developed a specific but simple to use OPCS coding table that listed all the breast and endocrine operations undertaken within the department along with their correct OPCS codes. The breast and endocrine consultant surgeons approved the coding table for its accuracy. The coding table was then distributed to all junior doctors within the breast and endocrine department, therefore making it accessible to doctors that complete the discharge summaries. A laminated copy of the coding table was posted in the operating theatre used by the breast and endocrine surgeons. An electronic version was uploaded in the shared hard drive of the breast and endocrine unit, and junior doctors were emailed an explanation of the coding table. To ensure durability, the coding table was added into the formal ‘’breast surgery junior doctors manual’ that junior doctors are provided with when they begin work in the breast and endocrine unit.

Strategy PDSA cycle 1: When designing the coding table and assigning OPCS codes to each operation, it became apparent that there was more than one appropriate OPCS code for certain operations. For example when coding a "breast wide local excision", OPCS codes B28.3 (“other excision, excision of lesion of breast”) or B28.2 (“other excision, partial excision of breast”) could be used. Following liaison with the hospital coding team, a meeting was held with the consultant breast and endocrine surgeons and a consensus was reached; for cancer excision operations the B28.2 code would be used, whereas the B28.3 code was more applicable for benign lump excisions such as fibroadenomas. PDSA cycle 2: The coding table was initially distributed between a few breast surgery junior doctors to test the design of the table. Feedback showed that the coding table made finding the correct OPCS code for a particular operation much easier and quicker than

before. It was suggested that the codes could be more logically listed, so the coding table was reconfigured and the amended coding table was arranged by operation type rather than in alphabetical order. PDSA cycle 3: Further feedback in the following weeks highlighted that some operations had the same OPCS code regardless of whether an additional procedure was performed as part of the primary operation. For example, when a breast wide local excision was performed the OPCS code was the same whether or not a sentinel node biopsy (SNB) procedure was also performed. This enabled the coding table to be further refined, for example by including +/- SNB after certain operations. PDSA cycle 4 and 5: The final version of the coding table (figure 1) was distributed among and approved by all members of the project team. The coding table was then published on the doctors’ shared computer hard drive and also in the written “breast surgery junior doctors manual”. In addition a laminated copy of the coding table was posted in the relevant operating theatres. Doctors using the coding table gave positive feedback regarding its use. In the first few weeks following the introduction of the final coding table, some doctors who were not involved in the coding table project were not aware of or were not using the table, so coding errors were still occurring. A short teaching session led by a representative from the hospital’s coding department was therefore organised for all members of the breast and endocrine team. The aim was to educate all staff on the importance of accurate OPCS coding and to highlight and encourage the use of the specifically-designed coding table. Following this, all doctors were using the coding table with noticeable benefits.

Results We prospectively re-audited for a two month period following implementation of the final coding table, and 113 operations were performed during this time. Of these, 73% were breast procedures, 16% were endocrine operations, and the remaining 11% comprised general surgical daycase procedures. A significant improvement was noted in the accuracy of OPCS procedural coding after the implementation of the coding table; 100% of the operations were correctly coded (p